Voice waveform generation method and device, server and storage medium

The embodiment of the invention provides a voice waveform generation method and device, a server and a storage medium. The method comprises the following steps: acquiring an input text; extracting condition features from the input text; and inputting the condition features into a waveform generation model obtained by training, and processing the condition features to obtain a voice waveform, wherein the waveform generation model comprises a prior distribution estimation network and a waveform generation network, the prior distribution estimation network is used for learning coding information of the natural voice waveform in a training stage, and the waveform generation network is used for generating the voice waveform according to the condition characteristics and an output result of the prior distribution estimation network. According to the embodiment of the invention, the waveform generation efficiency can be improved. 本申请实施例提供一种语音波形生成方法、装置、服务器及存储介质，该方法包括：获取输入文本；从输入文本中提取条件特征；将条件特征输入训练得到的波形生成模型，对条件特征进行处理，得